适配dtk23.04-km

README.md

@@ -77,3 +77,5 @@ hip结构体hipPointerAttribute_t适配
     src\array\cuda\gather_mm.cu:103
     src\array\cuda\gather_mm.cu:164
 ```
+3. 单元测试及python相关
+修改`np.int`、`np.float`、`np.asscalar`为`int`、`float`、`np.ndarray.item`，并将`setup.py`中`numpy`依赖版本提高至`1.20.0`，统一dtype相关类型的代码，以避免单元测试和使用中问题
\ No newline at end of file

examples/pytorch/bgnn/BGNN.py

@@ -287,9 +287,9 @@ class BGNNPredictor:
 
         # initialize for early stopping and metrics
         if metric_name in ['r2', 'accuracy']:
-            best_metric = [np.float('-inf')] * 3  # for train/val/test
+            best_metric = [float('-inf')] * 3  # for train/val/test
         else:
-            best_metric = [np.float('inf')] * 3  # for train/val/test
+            best_metric = [float('inf')] * 3  # for train/val/test
 
         best_val_epoch = 0
         epochs_since_last_best_metric = 0

examples/pytorch/ggnn/train_gc.py

@@ -56,7 +56,7 @@ def main(args):
                 labels = labels.data.numpy().tolist()
                 dev_preds += preds
                 dev_labels += labels
-        acc = np.equal(dev_labels, dev_preds).astype(np.float).tolist()
+        acc = np.equal(dev_labels, dev_preds).astype(float).tolist()
         acc = sum(acc) / len(acc)
         print(f"Epoch {epoch}, Dev acc {acc}")
 
@@ -76,7 +76,7 @@ def main(args):
                 labels = labels.data.numpy().tolist()
                 test_preds += preds
                 test_labels += labels
-        acc = np.equal(test_labels, test_preds).astype(np.float).tolist()
+        acc = np.equal(test_labels, test_preds).astype(float).tolist()
         acc = sum(acc) / len(acc)
         test_acc_list.append(acc)
 

examples/pytorch/ggnn/train_ns.py

@@ -53,7 +53,7 @@ def main(args):
                 labels = labels.data.numpy().tolist()
                 dev_preds += preds
                 dev_labels += labels
-        acc = np.equal(dev_labels, dev_preds).astype(np.float).tolist()
+        acc = np.equal(dev_labels, dev_preds).astype(float).tolist()
         acc = sum(acc) / len(acc)
         print(f"Epoch {epoch}, Dev acc {acc}")
 
@@ -73,7 +73,7 @@ def main(args):
                 labels = labels.data.numpy().tolist()
                 test_preds += preds
                 test_labels += labels
-        acc = np.equal(test_labels, test_preds).astype(np.float).tolist()
+        acc = np.equal(test_labels, test_preds).astype(float).tolist()
         acc = sum(acc) / len(acc)
         test_acc_list.append(acc)
 

examples/pytorch/hilander/utils/evaluate.py

@@ -36,7 +36,7 @@ def evaluate(gt_labels, pred_labels, metric='pairwise'):
     with Timer('evaluate with {}{}{}'.format(TextColors.FATAL, metric,
                                              TextColors.ENDC)):
         result = metric_func(gt_labels, pred_labels)
-    if isinstance(result, np.float):
+    if isinstance(result, float):
         print('{}{}: {:.4f}{}'.format(TextColors.OKGREEN, metric, result,
                                       TextColors.ENDC))
     else:

examples/pytorch/mvgrl/graph/dataset.py

@@ -53,7 +53,7 @@ def process(dataset):
         with open('{0}_node_attributes.txt'.format(prefix), 'r') as f:
             for line in f:
                 node_attrs.append(
-                    np.array([float(attr) for attr in re.split("[,\s]+", line.strip("\s\n")) if attr], dtype=np.float)
+                    np.array([float(attr) for attr in re.split("[,\s]+", line.strip("\s\n")) if attr], dtype=float)
                 )
     else:
         print('No node attributes')
@@ -113,7 +113,7 @@ def process(dataset):
                 f = np.zeros(max_deg + 1)
                 f[graph.degree[u[0]]] = 1.0
                 if 'label' in u[1]:
-                    f = np.concatenate((np.array(u[1]['label'], dtype=np.float), f))
+                    f = np.concatenate((np.array(u[1]['label'], dtype=float), f))
                 graph.nodes[u[0]]['feat'] = f
     return graphs, pprs
 

examples/pytorch/ogb/deepwalk/reading_data.py

@@ -188,7 +188,7 @@ class DeepwalkDataset:
             node_degree = self.G.out_degrees(self.valid_seeds).numpy()
             node_degree = np.power(node_degree, 0.75)
             node_degree /= np.sum(node_degree)
-            node_degree = np.array(node_degree * 1e8, dtype=np.int)
+            node_degree = np.array(node_degree * 1e8, dtype=int)
             self.neg_table = []
             
             for idx, node in enumerate(self.valid_seeds):

examples/pytorch/ogb/line/reading_data.py

@@ -184,7 +184,7 @@ class LineDataset:
             node_degree = self.G.out_degrees(self.valid_nodes).numpy()
             node_degree = np.power(node_degree, 0.75)
             node_degree /= np.sum(node_degree)
-            node_degree = np.array(node_degree * 1e8, dtype=np.int)
+            node_degree = np.array(node_degree * 1e8, dtype=int)
             self.neg_table = []
             
             for idx, node in enumerate(self.valid_nodes):

examples/pytorch/pointcloud/pct/ShapeNet.py

@@ -98,9 +98,9 @@ class ShapeNetDataset(Dataset):
         print('Loading data from split ' + self.mode)
         for fn in tqdm.tqdm(self.file_list, ascii=True):
             with open(fn) as f:
-                data = np.array([t.split('\n')[0].split(' ') for t in f.readlines()]).astype(np.float)
+                data = np.array([t.split('\n')[0].split(' ') for t in f.readlines()]).astype(float)
             data_list.append(data[:, 0:self.dim])
-            label_list.append(data[:, 6].astype(np.int))
+            label_list.append(data[:, 6].astype(int))
             category_list.append(shapenet.synset_dict[fn.split('/')[-2]])
         self.data = data_list
         self.label = label_list
@@ -122,6 +122,6 @@ class ShapeNetDataset(Dataset):
         cat = self.category[i]
         if self.mode == 'train':
             x = self.translate(x, size=self.dim)
-        x = x.astype(np.float)
-        y = y.astype(np.int)
+        x = x.astype(float)
+        y = y.astype(int)
         return x, y, cat

examples/pytorch/pointcloud/point_transformer/ShapeNet.py

@@ -98,9 +98,9 @@ class ShapeNetDataset(Dataset):
         print('Loading data from split ' + self.mode)
         for fn in tqdm.tqdm(self.file_list, ascii=True):
             with open(fn) as f:
-                data = np.array([t.split('\n')[0].split(' ') for t in f.readlines()]).astype(np.float)
+                data = np.array([t.split('\n')[0].split(' ') for t in f.readlines()]).astype(float)
             data_list.append(data[:, 0:self.dim])
-            label_list.append(data[:, 6].astype(np.int))
+            label_list.append(data[:, 6].astype(int))
             category_list.append(shapenet.synset_dict[fn.split('/')[-2]])
         self.data = data_list
         self.label = label_list
@@ -122,6 +122,6 @@ class ShapeNetDataset(Dataset):
         cat = self.category[i]
         if self.mode == 'train':
             x = self.translate(x, size=self.dim)
-        x = x.astype(np.float)
-        y = y.astype(np.int)
+        x = x.astype(float)
+        y = y.astype(int)
         return x, y, cat

examples/pytorch/pointcloud/pointnet/ShapeNet.py

@@ -98,9 +98,9 @@ class ShapeNetDataset(Dataset):
         print('Loading data from split ' + self.mode)
         for fn in tqdm.tqdm(self.file_list, ascii=True):
             with open(fn) as f:
-                data = np.array([t.split('\n')[0].split(' ') for t in f.readlines()]).astype(np.float)
+                data = np.array([t.split('\n')[0].split(' ') for t in f.readlines()]).astype(float)
             data_list.append(data[:, 0:self.dim])
-            label_list.append(data[:, 6].astype(np.int))
+            label_list.append(data[:, 6].astype(int))
             category_list.append(shapenet.synset_dict[fn.split('/')[-2]])
         self.data = data_list
         self.label = label_list
@@ -122,6 +122,6 @@ class ShapeNetDataset(Dataset):
         cat = self.category[i]
         if self.mode == 'train':
             x = self.translate(x, size=self.dim)
-        x = x.astype(np.float)
-        y = y.astype(np.int)
+        x = x.astype(float)
+        y = y.astype(int)
         return x, y, cat

python/dgl/contrib/data/knowledge_graph.py

@@ -417,7 +417,7 @@ def _load_data(dataset_str='aifb', dataset_path=None):
 
             # sort indices by destination
             edge_list = sorted(edge_list, key=lambda x: (x[1], x[0], x[2]))
-            edge_list = np.asarray(edge_list, dtype=np.int)
+            edge_list = np.asarray(edge_list, dtype=int)
             print('Number of edges: ', len(edge_list))
 
             np.savez(edge_file, edges=edge_list, n=np.asarray(num_node), nrel=np.asarray(num_rel))

python/dgl/data/minigc.py

@@ -156,7 +156,7 @@ class MiniGCDataset(DGLDataset):
         for i in range(self.num_graphs):
             # convert to DGLGraph, and add self loops
             self.graphs[i] = add_self_loop(from_networkx(self.graphs[i]))
-        self.labels = F.tensor(np.array(self.labels).astype(np.int))
+        self.labels = F.tensor(np.array(self.labels).astype(int))
 
     def _gen_cycle(self, n):
         for _ in range(n):

python/setup.py

@@ -168,7 +168,7 @@ setup(
     maintainer_email='wmjlyjemaine@gmail.com',
     packages=find_packages(),
     install_requires=[
-        'numpy>=1.14.0',
+        'numpy>=1.20.0',
         'scipy>=1.1.0',
         'networkx>=2.1',
         'requests>=2.19.0',

tests/compute/test_data.py

@@ -510,12 +510,12 @@ def _test_construct_graphs_multiple():
     num_edges = 1000
     num_graphs = 10
     num_dims = 3
-    node_ids = np.array([], dtype=np.int)
-    src_ids = np.array([], dtype=np.int)
-    dst_ids = np.array([], dtype=np.int)
-    ngraph_ids = np.array([], dtype=np.int)
-    egraph_ids = np.array([], dtype=np.int)
-    u_indices = np.array([], dtype=np.int)
+    node_ids = np.array([], dtype=int)
+    src_ids = np.array([], dtype=int)
+    dst_ids = np.array([], dtype=int)
+    ngraph_ids = np.array([], dtype=int)
+    egraph_ids = np.array([], dtype=int)
+    u_indices = np.array([], dtype=int)
     for i in range(num_graphs):
         l_node_ids = np.random.choice(
             np.arange(num_nodes*2), size=num_nodes, replace=False)
@@ -1191,7 +1191,7 @@ def _test_NodeEdgeGraphData():
     from dgl.data.csv_dataset_base import NodeData, EdgeData, GraphData
     # NodeData basics
     num_nodes = 100
-    node_ids = np.arange(num_nodes, dtype=np.float)
+    node_ids = np.arange(num_nodes, dtype=float)
     ndata = NodeData(node_ids, {})
     assert np.array_equal(ndata.id, node_ids)
     assert len(ndata.data) == 0
@@ -1228,8 +1228,8 @@ def _test_NodeEdgeGraphData():
     assert len(edata.data) == 0
     assert np.array_equal(edata.graph_id, np.full(num_edges, 0))
     # EdageData more
-    src_ids = np.random.randint(num_nodes, size=num_edges).astype(np.float)
-    dst_ids = np.random.randint(num_nodes, size=num_edges).astype(np.float)
+    src_ids = np.random.randint(num_nodes, size=num_edges).astype(float)
+    dst_ids = np.random.randint(num_nodes, size=num_edges).astype(float)
     data = {'feat': np.random.rand(num_edges, 3)}
     etype = ('user', 'like', 'item')
     graph_ids = np.arange(num_edges)
@@ -1259,7 +1259,7 @@ def _test_NodeEdgeGraphData():
     assert np.array_equal(gdata.graph_id, graph_ids)
     assert len(gdata.data) == 0
     # GraphData more
-    graph_ids = np.arange(num_graphs).astype(np.float)
+    graph_ids = np.arange(num_graphs).astype(float)
     data = {'feat': np.random.rand(num_graphs, 3)}
     gdata = GraphData(graph_ids, data)
     assert np.array_equal(gdata.graph_id, graph_ids)

tests/compute/test_heterograph.py

@@ -1125,8 +1125,8 @@ def test_convert(idtype):
         dsttype = hg.ntypes[ntype_id[dst[i]]]
         etype = hg.etypes[etype_id[i]]
         src_i, dst_i = hg.find_edges([eid[i]], (srctype, etype, dsttype))
-        assert np.asscalar(F.asnumpy(src_i)) == nid[src[i]]
-        assert np.asscalar(F.asnumpy(dst_i)) == nid[dst[i]]
+        assert np.ndarray.item(F.asnumpy(src_i)) == nid[src[i]]
+        assert np.ndarray.item(F.asnumpy(dst_i)) == nid[dst[i]]
 
     mg = nx.MultiDiGraph([
         ('user', 'user', 'follows'),